Optimization of the Arterial Spin Labeling Input function (ASL-IF) to stabilize the perfusion signal

• Applicant: Dr. Thomas Lindner
• Subject Area: Radiology ; Clinical Neurology; Neurosurgery and Neuroradiology ; Medical Physics, Biomedical Technology
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 513280552

Perfusion imaging is vital for assessing blood flow in organs like the brain, aiding in the early detection of conditions such as stroke or tumors. MRI-based techniques like dynamic contrast-enhanced MRI (DCE-MRI) use gadolinium contrast agents to measure perfusion, but concerns about gadolinium deposition and risks in renal failure patients have increased interest in non-invasive alternatives. Arterial spin labeling (ASL) uses magnetically labeled blood as an endogenous tracer to avoid external contrast agents. Pseudo-continuous ASL (PCASL) is the currently recommended technique to quantify cerebral blood flow (CBF), though its accuracy depends on the subject-specific labeling efficiency (LE), which is often approximated with fixed values, potentially leading to errors in pathological cases. In a prior project PCASL was enhanced by optimizing measurements for patients and exploring its diagnostic potential in glioma subtyping and carotid artery stenosis. These studies showed ASL’s ability to differentiate glioma grades via histogram and radiomics analyses, though it struggled with specific molecular markers. In carotid stenosis, PCASL reliably identified affected sides, with multi-PLD and single-PLD approaches showing equivalent diagnostic value. A key focus was improving the LE estimation using the Arterial Spin Labeled Input Function (ASLIF), which monitors labeled blood during PCASL labelling. In this project, quantitative ASLIF model to estimate LE, flow velocity, and blood relaxation time (T1b) was developed and then validated with flow phantoms and initial in-vivo tests. Current limitations include ASLIF’s one-dimensional imaging, restricting artery-specific data and its reliance on early signal phases, limiting T1b estimation. Data processing is also offline, delaying clinical use. Thus, remaining challenges involve refining the ASLIF model to handle dispersion effects, improving robustness against cardiac and field variations, enabling two-dimensional imaging, and achieving real-time analysis. This project aims to advance ASLIF and PCASL for reliable, patient-tailored perfusion imaging in clinical practice.

• DFG Programme: Research Grants